As complex workloads that run on many servers are pursuing higher networking throughput, more CPU cycles are consumed to support the TCP stack. To mitigate the high CPU burden from executing the compute-intensive TCP, prior works have proposed to offload TCP processing to the embedded processors, ASICs, or FPGAs in network devices. However, none of the approaches satisfy all of the critical requirements of TCP simultaneously, which are high performance, many connections, and high flexibility. Embedded processors do not provide enough performance to fully offload the TCP stack, while ASICs fail to provide enough flexibility. Meanwhile, existing FPGA-based TCP accelerators either fail to provide high performance or give up some of the critical features and requirements to achieve high performance due to their inefficient processing architecture.

Cryogenic computing, which runs a computer device at an extremely low temperature, is promising thanks to its significant reduction of wire resistance as well as leakage current. Recent studies on cryogenic computing have focused on various architectural units including the main memory, cache, and CPU core running at 77K. However, little research has been conducted to fully exploit the fast cryogenic wires, even though the slow wires are becoming more serious performance bottleneck in modern processors. In this paper, we propose a CPU microarchitecture which extensively exploits the fast wires at 77K. For this goal, we first introduce our validated cryogenic-performance models for the CPU pipeline and network on chip (NoC), whose performance can be significantly limited by the slow wires. Next, based on the analysis with the models, we architect CryoSP and CryoBus as our pipeline and NoC designs to fully exploit the fast wires. Our evaluation shows that our cryogenic computer equipped with both microarchitectures achieves 3.82 times higher system-level performance compared to the conventional computer system thanks to the 96% higher clock frequency of CryoSP and five times lower NoC latency of CryoBus.

To understand and mimic the working mechanism of the brain, neuroscientists rely on brain simulations that operate in a time-driven manner. The simulation involves evaluating how the neurons change their states over time and transferring spikes to the connected neurons through synapses. It also simulates learning by evaluating how the synapses change their weights according to the spiking activity of the neurons. To explore various behaviors of the brain and thus make great advances, neuroscientists need a methodology to support large-scale simulations in both an accurate and efficient manner. For accurate simulations, existing simulators adopt a time-precise simulation methodology where the simulator computes all the neuronal and the synaptic state changes in time order. Unfortunately, they suffer from significant underutilization and energy inefficiency as the simulator scales.In this paper, we present NeuroSync, a fast, energy-efficient, and scalable hardware-based accelerator for accurate brain simulations. The key idea is to adopt a speculative simulation methodology at a minimum overhead along with architectural support. NeuroSync achieves high efficiency using an optimal dataflow for the speculative simulations. At the same time, it ensures simulation accuracy by carefully designing a rollback and recovery mechanism to handle mis-speculations. To implement the methodology at a low cost, NeuroSync further proposes a speculation-optimal learning simulation method. Our evaluations show that 64-chip NeuroSync achieves 3.37× speedup and 3.81× higher energy efficiency with only 10.96% area overhead. The evaluations also show that NeuroSync is extremely scalable with higher speedup as the system scales.

Brain-inspired computing aims to understand the cognitive mechanisms of a brain and apply them to advance various areas in computer science. Deep learning is an example to greatly improve the field of pattern recognition and classification by utilizing an artificial neural network (ANN). To exploit advanced mechanisms of a brain and thus make more great advances, researchers need a methodology that can simulate neural networks with higher computational capabilities such as advanced spiking neural networks (SNNs) with two-stage neurons and synaptic delays. However, existing SNN simulation methodologies are too slow and energy-inefficient due to their software-based simulation or hardware-based but time-driven execution mechanisms.

MOTIVATION: The identification and characterization of genes that increase the susceptibility to common complex multifactorial diseases is a challenging task in genetic association studies. The multifactor dimensionality reduction (MDR) method has been proposed and implemented by Ritchie et al. (2001) to identify the combinations of multilocus genotypes and discrete environmental factors that are associated with a particular disease. However, the original MDR method classifies the combination of multilocus genotypes into high-risk and low-risk groups in an ad hoc manner based on a simple comparison of the ratios of the number of cases and controls. Hence, the MDR approach is prone to false positive and negative errors when the ratio of the number of cases and controls in a combination of genotypes is similar to that in the entire data, or when both the number of cases and controls is small. Hence, we propose the odds ratio based multifactor dimensionality reduction (OR MDR) method that uses the odds ratio as a new quantitative measure of disease risk. RESULTS: While the original MDR method provides a simple binary measure of risk, the OR MDR method provides not only the odds ratio as a quantitative measure of risk but also the ordering of the multilocus combinations from the highest risk to lowest risk groups. Furthermore, the OR MDR method provides a confidence interval for the odds ratio for each multilocus combination, which is extremely informative in judging its importance as a risk factor. The proposed OR MDR method is illustrated using the dataset obtained from the CDC Chronic Fatigue Syndrome Research Group. AVAILABILITY: The program written in R is available.

MOTIVATION: The identification and characterization of susceptibility genes that influence the risk of common and complex diseases remains a statistical and computational challenge in genetic association studies. This is partly because the effect of any single genetic variant for a common and complex disease may be dependent on other genetic variants (gene-gene interaction) and environmental factors (gene-environment interaction). To address this problem, the multifactor dimensionality reduction (MDR) method has been proposed by Ritchie et al. to detect gene-gene interactions or gene-environment interactions. The MDR method identifies polymorphism combinations associated with the common and complex multifactorial diseases by collapsing high-dimensional genetic factors into a single dimension. That is, the MDR method classifies the combination of multilocus genotypes into high-risk and low-risk groups based on a comparison of the ratios of the numbers of cases and controls. When a high-order interaction model is considered with multi-dimensional factors, however, there may be many sparse or empty cells in the contingency tables. The MDR method cannot classify an empty cell as high risk or low risk and leaves it as undetermined. RESULTS: In this article, we propose the log-linear model-based multifactor dimensionality reduction (LM MDR) method to improve the MDR in classifying sparse or empty cells. The LM MDR method estimates frequencies for empty cells from a parsimonious log-linear model so that they can be assigned to high-and low-risk groups. In addition, LM MDR includes MDR as a special case when the saturated log-linear model is fitted. Simulation studies show that the LM MDR method has greater power and smaller error rates than the MDR method. The LM MDR method is also compared with the MDR method using as an example sporadic Alzheimer's disease.